Skip to contents

Score a gorilla questionnaire using a survey key

Source: R/gorilla_q_keyed.R
gorilla_q_keyed.Rd

Processes questionnaires made with Questionnaire Builder 1 (long format data). A survey key is a .csv file that tells this function how to score each item, which items to reverse, and which items belong to which scale or subscale. The result has one row per participant.

Usage

gorilla_q_keyed(
  data,
  keyout = FALSE,
  qlist = NULL,
  items = F,
  key_folder = NULL
)

Arguments

data

The list of data from a gorilla import, or just the questionnaire data (data_q). It has to be in long format.

keyout

Set to TRUE to write out a new blank key for each questionnaire, from the data. This replaces any existing key in survey_key.

qlist

The names of the questionnaires (Task.Name in gorilla) to score. If NULL, all of the questionnaires in the data.

items

If TRUE, also return the individual item scores (after reversing etc.), one column for each item.

key_folder

Folder where the survey keys are kept. If NULL, the folder survey_key in the working directory is used.

Value

A data.table with one row per participant (pid) and a column for each scale (and subscale), and for any items that were left out of scoring. If nothing could be scored, there is no return value.

Details

How to use it

  1. The first time, run with keyout = TRUE. This writes a blank key for each questionnaire in qlist to a folder called survey_key in your working directory.

  2. Edit the key in Excel or a text editor, and save it as a .csv file. There is one key per questionnaire in gorilla, named after the questionnaire.

  3. Run the function again (with keyout = FALSE). It looks for the key in key_folder, or if that is not given in the folder survey_key.

Each row of the key is one item, identified by the Question.Key (what you typed in the "key" box when you made the questionnaire in gorilla). The columns you can edit are:

  • sum - how the item is used in the score: 1 to add it, -1 to reverse score it, and 0 to leave it out of the scores (it is returned by itself in its own column).

  • rev - for reverse scored items, the answer is subtracted from this number. eg for a 7 point scale, use 8 (the default).

  • adj - a number added to the item, if the scale needs adjusting (eg it starts at 0 instead of 1). Optional: keys without this column are treated as 0.

  • scaleName - the name of the scale that the item adds to. The output has a column with this name. One questionnaire can have many scales.

  • Subscore - optional name of a subscale. It is returned as an extra column <scaleName>_<Subscore>.

  • qual - put 1 here if the item is a text (qualitative) answer. It is not scored but returned in its own column.

  • ignore - an answer that should be ignored when scoring, eg 8 for "not applicable" on a scale of 1 to 7.

Scales are the sum of the items (not the mean). People who did not answer all the items are listed.

Questionnaires without a key are skipped, unless you make one with keyout = TRUE. If you only want to look at the data without scoring it, see gorilla_q_parse().

See also

gorilla_q_keyed_qb2(), gorilla_q_parse()

Examples

if (FALSE) { # \dontrun{
# make a blank key
gorilla_q_keyed(data, qlist = "Interpersonal Reactivity Index (Davies, 1983)", keyout = TRUE)
# ... edit the key in survey_key, then score:
gorilla_q_keyed(data, qlist = "Interpersonal Reactivity Index (Davies, 1983)")
} # }